Create a simplified combined groundwater deviation map per vegetation

[cecileherr]

• niche_vlaanderen version: > 1.1 (not implemented yet)
• Python version: 3

Niche Python Package

It would be useful to create a simplified version of the groundwater deviation maps combining the (interpreted) deviation for mean highest (mhw_x) and mean lowest groundwater levels (mlw_x) in a few descriptive categories for each vegetation type x.
For each vegetation type, there would be one extra raster then with all the (theoretically) possible combinations of the qualitative classification of the mean highest and lowest groundwater levels. The legend would be:

mhw	mlw
suitable	suitable
suitable	too wet
suitable	too dry
too wet	suitable
too wet	too wet
too wet	too dry
too dry	suitable
too dry	too wet
too dry	too dry

Suitable means within the range (min-max) provided in the NICHE reference table, and is reflected by a value 0 in the respective deviations maps. Too wet means a negative, and too dry a positive value in the respective deviation maps.

The resulting raster could probably be created and exported automatically along with the deviation maps (to avoid having to use another option). deviation=True would then create the 28*2 (quantitative) deviation maps + an extra set of 28 maps with the result of the (qualitative) combinations of the classified versions of these rasters for each vegetation type.

The (color) coding for the different combinations should be constant even if some categories are missing.

[johanvdw]

https://colorbrewer2.org/#type=qualitative&scheme=Paired&n=9

[stijnvanhoey]

Idea of the legend/color-scheme:

The name is "Bivariate Choropleth Maps", e.g. https://www.joshuastevens.net/cartography/make-a-bivariate-choropleth-map/. A Python/matplotlib version is shown inhttps://waterprogramming.wordpress.com/2022/09/08/bivariate-choropleth-maps/

[stijnvanhoey]

Note that the bivariate colormap is expected to be included into matplotlib directly in October 2024, see blog post on summer of code